Model selection and signal extraction using Gaussian Process regression

We present a novel computational approach for extracting weak signals, whose exact location and width may be unknown, from complex background distributions with an arbitrary functional form. focus on datasets that can naturally presented as binned integer counts, demonstrating our the CERN open dataset ATLAS collaboration at Large Hadron Collider, which contains Higgs boson signature. Our is based Gaussian Process (GP) regression - powerful flexible machine learning technique allowed us to model without specifying its form explicitly, separate signal contributions in robust reproducible manner. Unlike fits, GP-regression-based does not need constantly updated more data becomes available. discuss how select GP kernel type, considering trade-offs between complexity ability capture features of distribution. show framework used detect resonance statistical significance than polynomial fit specifically tailored dataset. Finally, we use Markov Chain Monte Carlo (MCMC) sampling confirm extracted